Suppression of RBC alloimmunization and regulation of CD4⁺T cell dependence by C3 is not due to genetic confounders in mice

Abstract

Background

Activation of complement protein C3 generally enhances antibody responses and C3‐null mice have decreased antibody‐based immunity. Mener et al. have reported a paradoxical suppressor function for C3 in alloimmunization to transfused RBCs as alloantibodies are increased in C3‐null mice. Moreover, C3 regulated the CD4⁺ T cell dependence of the immune response. However, the C3‐null mice used were on a mixed B6/129 genetic background. We have previously reported that 129 mice have significantly higher alloimmune responses to RBC transfusion and we mapped a genetic locus that contains C3 (amongst other genes). Given the surprising nature of Mener et al.'s findings and the potential confounding from 129 genetic elements, it is a necessary part of scientific rigor to suspect that contaminating 129 elements rather than the deletion of C3 caused the observed biology.

Methods

We used CRISPR/Cas9 to generate a new C3‐null mouse (C3Cr‐KO) directly in B6 mice lacking any 129 genetic elements. B6 and C3Cr‐KO mice were transfused with KEL‐K2_med RBCs with or without CD4⁺ T cell depletion and serum α‐KEL IgM and Igs were quantified by crossmatch.

Results

Identical to the findings of Mener et al., alloimmunization was increased in C3Cr‐KO mice compared to wild‐type B6 mice and CD4⁺T cell dependence of the alloimmune response was reversed.

Conclusions

The current findings eliminate a common confounder present in murine knockout systems that has caused erroneous conclusions in other settings. Both the conclusion that the presence of the C3 gene decreases RBC alloimmunization and regulates CD4⁺ T cell dependence was confirmed.

Abbreviations

C3

Complement 3

C3 KO

Complement 3 knock out

C3Cr‐KO

Complement 3 CrispR knock out

CR1/CR2/CR3/CR4

Complement Receptor1/2//3/4

ES

Embryonic Stem

HOD

Hel Ova Duffy

KO

Knock Out

MFI

Mean Fluorescence Intensity

RBC

Red Blood Cells

SNP

Single Nucleotide Polymorphism

1. INTRODUCTION

Alloimmunization to transfused red blood cells (RBCs) remains a substantial barrier to ongoing therapy in chronically transfused patients who become alloimmunized to multiple blood group antigens. ¹ Progress in recent decades has also demonstrated that the mechanisms of alloimmunization to transfused RBCs are distinct from other better‐studied immune stimuli (e.g., microbial infections, vaccines, etc.). Alloimmunization to RBCs can induce immunity or tolerance depending upon the inflammatory status of the recipient at the time of transfusion. ² , ³ , ⁴ , ⁵ , ⁶ In conditions where alloimmunization to RBCs occurs, a spleen is required due to the functions of two cell types only found in the splenic environment (i.e., bridging channel dendritic cells and also marginal zone B cells). ⁷ , ⁸ , ⁹ , ¹⁰ In addition, type I interferons, ¹¹ , ¹² interleukin 6, ¹³ and CD40 ¹⁴ are also required and/or involved in RBC alloimmunization. RBC alloantigen copy number is a critical determinant and can regulate immunity versus tolerance as well as affecting CD4⁺T cell dependence or independence of an alloantibody response. ¹⁵ , ¹⁶ Both the role of inflammation and the requirement for a spleen, first discovered in mice, have translated into human biology. ⁵ , ⁶ , ¹⁷ , ¹⁸

We and others have reported a complex and central role for complement C3 protein in regulating RBC alloimmunization. ¹⁹ , ²⁰ , ²¹ , ²² Cleavage of C3 by the complement cascade is generally associated with activation of innate immunity and enhancement of adaptive immunity in multiple settings. ²³ , ²⁴ In contrast, Mener et al. reported a surprising finding that RBC alloimmunization is enhanced in C3 knockout mice suggesting that C3 suppresses rather than enhances RBC alloimmunization. ²⁰ While a negative effect of C3 on adaptive immunity is not unprecedented ²⁴ it is not typical during primary humoral immunity. In a subsequent report, Mener et al. also demonstrated that RBC alloimmunization that is CD4⁺ T cell independent becomes CD4⁺ T cell dependent in C3 knockout mice. ²²

Both of the above reports by Mener et al. represent seminal findings that were unanticipated, given the general effects of C3 in other settings. ²⁰ , ²² Importantly, both the conclusion that C3 suppresses RBC alloimmunization and that C3 regulates CD4⁺ T cell dependence were drawn from the altered immune responses of C3 knockout mice. The mechanistic interpretation of knockout mouse findings requires at least three conditions to be valid. First, the targeted gene must actually be knocked out. Second, no off‐target effects can be present. Third, the knockout mouse can differ from the wild‐type control only with regard to the knocked‐out gene. Unless such criteria are filled, causal inference for a mechanistic role of the knocked‐out gene is susceptible to multiple confounders.

Historically, the majority of knockout mice have been generated using embryonic stem (ES) cells derived from substrains of 129 mice. The knockout mice are then backcrossed for multiple generations with C57BL/6 (B6) mice allowing the use of wild‐type B6 mice as the control strain. However, there is inevitably contamination of 129 genetic elements in the backcrossed knockout mice. ²⁵ Contamination can occur from incomplete backcrossing but is also inescapable for any genes in close proximity to the knockout gene that are co‐inherited during the backcrossing process. ²⁶ Indeed, there are well‐documented cases of biological effects being misattributed to knocked‐out genes that were actually caused by 129 variants in close proximity. ²⁷ , ²⁸

The only commonly available C3 knockout mice (C3 KO), which were used for all published studies on RBC alloimmunization, were generated using 129S4/SvJae derived ES cells. ²⁹ Moreover, the very provocative studies of C3 effects on RBC alloimmunization utilized mice that were only partially backcrossed onto a B6 backgound. ²² , ³⁰ Finally, the C3 KO mouse was made at a time when antibiotic selection cassettes were integrated into the genome at the knockout site, and these cassettes are known to affect the expression of third‐party genes close to the knocked‐out gene. ³¹

The unanticipated findings that C3 suppresses RBC alloimmunization and regulates CD4⁺ T cell dependence were somewhat extraordinary, and as popularized by Carl Sagan, in science “extraordinary claims require extraordinary evidence.” It is thus reasonable to be suspicious that the findings of C3 effects on RBC alloimmunization were an artifact of 129 genetic elements and/or the neomycin cassette that was left behind.

The suspicion about 129 genetic elements was of particular concern in light of our previous report that 129 mice have a significantly higher alloimmune response to transfused RBCs ³² and that the C3 KO mice used by Mener et al. have substantial contamination from 129 mice on 13 out of 21 chromosomes. ²⁵ Knocking out the CR1/CR2 complement receptors had the same effect on CD4⁺ T cell dependence as C3 KO mice, consistent with a role of C3 itself. In contrast, confidence in concluding C3 suppresses alloimmunization was decreased by the observation that CR1/CR2 KO mice did not have increased RBC alloimmunization, as was seen in the C3 KO mouse, but rather decreased alloimmunization consistent with the normal adjuvant effect of C3 activation. ³⁰

For each of the above reasons, we hypothesized that the altered alloimmune responses in C3 knockout mice were due not to the deletion of C3 but rather to a genetic contaminant from 129 mice, causing a mistaken interpretation of C3 effects.

2. MATERIALS AND METHODS

2.1. Mice

Fertilized eggs from C57BL/6J mice were electroporated with recombinant CAS9 (IDT [Coralville, IA]) and a guide RNA targeted to exon 4 of the C3 gene (GACAACAACCTACTGCCCGT). The guide RNA was generated by using the CRISPR guide design algorithm CRISPOR (http://crispor.tefor.net/). Zygotes that progressed to the two‐cell stage were implanted in the oviducts of pseudo pregnant foster mothers. Pups born from the foster mothers were screened using tail snip DNA by PCR genotyping of exon 4. A founder with the indicated mutation was isolated and germline transmission was confirmed by breeding the founder with wild‐type C57BL/6J mice. The newly generated animal is designated as C3Cr‐KO. Wild‐type C57BL/6J mice and C3 KO mice either on a mixed B6.129 background (B6.129S4‐C3 ^tm1Crr /JJAX:003641) or a B6 background (B6.129S4‐C3 ^tm1Crr /J;JAX 029661) were purchased from Jackson Laboratories (Bar Harbor, ME). KEL‐K2_med and KEL‐K2_hi mice were bred in a vivarium at the University of Virginia. All procedures were carried out under approved protocols from the University of Virginia Animal Care and Use Committee (ACUC).

2.2. Complement C3 ELISA and fixation assay

Serum was harvested from B6, C3KO (Jackson lab) and C3Cr‐KO mice. The activity of Complement C3 in the serum was detected by the deposition of C3 on B6 RBCs coated with rabbit polyclonal mouse RBC antibody (Rockland) and was measured by an anti‐C3 antibody that recognizes all C3 forms (Cedarlane Labs, biotinylated clone RmC11H9). Complement C3 level in the serum was measured using the Mouse C3 ELISA kit (Abcam) as per the manufacturer's directions.

2.3. Depletion and measurement of CD4 ⁺ T cells

Mice were treated with three intraperitoneal doses of 250 μg of either monoclonal anti–mouse CD4 (clone GK1.5, Bio X Cell) or isotype‐matched control IgG2b (clone LTF2, Bio X Cell) at days −4 and −2 days prior to transfusion and day 7 post transfusion. Depletion was confirmed by staining peripheral blood with anti–mouse CD4 antibody (clone RM4, BioLegend) and analyzing by flow cytometry.

2.4. Transfusion and measurement of RBC alloimmunization

Mice were transfused with 50 μL of packed KEL‐K2_med RBCs via lateral tail vein injections. Peripheral blood was obtained by retro‐orbital bleeds at the indicated time points. A flow cytometric crossmatch assay was used to measure alloantibody responses in transfusion recipients using fluorescently labeled secondary antibodies specific for either IgM or total IgGs (Southern Biotech). In brief, sera collected at the indicated time points were incubated with target RBCs expressing either KEL‐K2_hi B6 RBCs followed by incubation with the fluorescently labeled secondary antibody. The antigen‐specific response (i.e., adjusted MFI) was determined by subtracting the MFI of B6 RBCs (background) from the MFI of KEL‐K2_hi RBC targets. KEL‐K2_hi RBCs were used as targets because the higher levels of KEL‐K2 expression result in more sensitive antibody detection. Attune NxT Cytometer (Thermo Fisher Scientific) and FlowJo software were used for data collection and analysis.

2.5. Statistical analysis

Statistical analysis was performed using a two‐way ANOVA with Šidák's multiple‐comparison test. p values of less than .05 were considered statistically significant. All data are presented as the mean ± SEM.

3. RESULTS

3.1. Multiple immune regulatory genes are in linkage disequilibrium with the C3 gene

A version of the C3 KO mouse used by Mener et al. more extensively backcrossed onto a B6 background over many generations was generated by Jackson Labs precisely to address issues of 129 genetic contamination (referred to as Jax 029661). Analysis of these mice showed the identical phenotype to that reported by Mener et al. In particular the C3 KO mice had increased alloantibody responses and a CD4⁺ dependence to transfused RBCs compared to wild‐type B6 mice (Figure S1). Although these data strengthened confidence in the conclusion that C3 was responsible, there have been published examples of congenic mice where the effects were not due to knocking out a gene but were rather due to other 129 genes that were coinherited during congenesis with the knocked‐out gene due to close proximity in the genome.

To more thoroughly characterize the precise genetics of the congenic C3 KO mouse, DNA was analyzed with a panel of single nucleic acid polymorphisms (SNPs) that were informative between 129 and B6 strains (Figure 1A). As predicted, the only detected 129 genetic elements were flanking the C3 gene and were homozygous in all mice tested (n = 4). The donated 129 genome ranged from 52,611,561 to 61,541,220 base pairs and contained 84 protein coding genes, 89 long noncoding RNA genes (lncRNA) 14 other non‐coding RNA based genes, and 64 pseudogenes (Table S1). Other than C3, 8 of the coding genes are directly involved in humoral immunity either through direct effects on B cell biology (Vav1, CD70, and Stap3), costimulation of helper T cells (4‐1BBL), regulating T‐B cell interactions (Ebi3 and LIGHT), effects on antigen presenting cells (Dennd1c), or regulatory effects on the immune environment (Mydgf). In addition, one of the non‐coding RNA genes (miR‐7b) has been reported to affect CD4⁺ T cell biology. ³³ Thus, the more extensively backcrossed congenic mouse does not eliminate confounders with the potential to alter antibody responses and remained a concern, especially for the enhancement of RBC alloimmunization in C3 KO mice that was not recapitulated in CR1/CR2 KO mice. Regrettably, due to extremely low frequencies of recombination in small congenic regions, further isolation of C3 by backcrossing was not feasible.

FIGURE 1.

Generation of C3Cr‐KO mice by targeted deletion of the C3 gene using CRISPR/Cas9. (A) Mouse complement C3 exon 4 in B6 and C3Cr‐KO. Guide RNA targeted to exon 4 resulted in a prematurely truncated C3 protein. (B) Serum isolated from B6, C3KO (Jax 029661), and C3Cr‐KO were assayed by ELISA to detect the presence of complement C3. (C) C3 functional activity was measured in the serum from B6, C3KO(Jackson) and C3Cr‐KO by complement C3 fixation assay.

3.2. Generation and characterization of C3 knockout mouse on B6 background

Fertilized eggs from C57BL/6J mice were microinjected with CAS9 enzyme and guide RNAs targeting exon 4 of the C3 gene (see methods for details). A single founder was isolated that had an 8 nucleotide deletion in exon 4 resulting in a predicted gene product consisting of the first 158 amino acids out of the 1664 amino acids of the C3 protein followed by a frame shift resulting in an aberrant 7 amino acids (QDSRHPH) before hitting a stop codon (Figure 1A). The murine C3 gene has 41 exons with exons 1–16 encoding the beta chain and exons 16–41 encoding the alpha chain. All active sites for C3 are down stream of exon 4. As such, the observed deletion is predicted to eliminate all known C3 function. This mouse is referred to as C3Cr‐KO for the rest of this manuscript.

Serum from C3Cr‐KO mice was tested for the presence of C3 protein and activity using B6 serum as a positive control and serum from C3 KO mice as a negative control. Both C3 protein and C3 activity were easily detected in B6 but not C3 KO serum. (Figure 1B,C) Similar to C3 KO mice, neither C3 protein nor C3 activity was detected in C3Cr‐KO serum by ELISA (Figure 1B) or the RBC complement fixation assay, respectively (Figure 1C). Together, the above data demonstrate the generation of a novel C3 null mouse that is isogenic to C57BL/6J mice with neither 129 genetic contaminants nor a residual NEO cassette as found in the commercially available C3 KO mouse.

3.3. C3cr‐KO mice have increased alloimmunization and a CD4 ⁺ T cell dependent response to KEL‐K2_med

Male and female C3cr‐KO mice were treated with either a CD4⁺ depleting antibody (GK1.5) or an isotype control (IgG2b) followed by a transfusion with KEL‐K2_med RBCs and alloantibodies to KEL‐K2 were assayed at days 7,14, and 21 post‐transfusion. (Figure 2A). The effectiveness of CD4⁺ T cell depletion was confirmed by staining peripheral blood from these mice at day 7 (Figure 2B,C).

FIGURE 2.

GK1.5 treatment depletes CD4⁺T cells in the peripheral blood. (A) Experimental scheme depicting CD4⁺T cell depletion, transfusion and phlebotomy schedule. The figure was generated on Biorender.com (B) Peripheral blood from mice treated either with CD4⁺T cell depleting antibody GK1.5 or matched isotype control IgG2b antibody was assessed to confirm CD4⁺T cell absence.

Consistent with the findings in C3 KO mice from the previous reports, anti‐KEL‐K2 IgG was higher in C3cr‐KO mice than in wild‐type controls (Figure 3A–F). Also consistent with the previous report, depletion of CD4⁺ T cells had no significant effect in wild‐type B6 recipients but essentially eliminated alloimmunization in C3cr‐KO mice. As recipient gender is also a potential variable in allo‐immune response. ³⁴ Thus, we transfused both male (Figure 3A–C) and female (Figure 3 D‐F) B6 and C3Cr‐KO recipients with KEL‐K2med in the presence or absence of CD4+T cell depletion and observed no sex‐specific alteration of immune response. There were no significant changes in early IgM responses (Figure 3A,D). There was no difference in mice receiving isotype control compared to control untreated B6 mice (data not shown).

FIGURE 3.

Alloimmune response to KEL‐K2med is CD4⁺T cell dependent in the absence of C3 both in male and female mice. (A–F) Mice were bled at the indicated time points and amount of anti‐KEL IgM at day 7 and anti‐KEL Igs were measured by a flow cytometry based cross match assay using either KEL‐K2_hi or B6 RBCs as targets as detailed in Section 2. Adjusted MFIs are calculated by subtracting the background antibody signal on antigen‐negative B6 RBCs from RBCs expressing KEL‐K2_hi. Representative experiments with at least 3 repeats with similar outcomes with 5 mice per group are shown. p values were calculated using a repeated‐measures, two‐way ANOVA with Šidák's multiple‐comparisons test and are designated as p > .05 (NS) and *p < .05, **p < .01, and ***p < .0005.

4. DISCUSSION

The current report confirms that deletion of C3 both enhances alloimmunization to KEL‐K2_med and converts the alloimmune response from CD4⁺ T cell independent to CD4⁺ dependent. The findings by Mener et al. have substantially forwarded the field especially in that they demonstrate a role for C3 that is generally inconsistent with known C3 biology in other settings. ²² , ³⁰ It is precisely because these findings are highly novel that they require a greater level of scrutiny, especially as we have documented that 129 mice have increased alloimmunization to RBC transfusion and mapped a QTL that contains the C3 gene. ³² The findings in the current manuscript provide the additional rigor required in light of the surprising findings made by Mener et al. to rule out a known confounder in murine knockout experiments (i.e., effects due to genetic elements inherited from ES cells used to make KO mice and/or effects of residual NEO cassettes). ²⁷ , ²⁸

In the strictest interpretation, neither the data in this manuscript nor the findings from Mener et al. ²² , ³⁰ demonstrate a direct role for C3 in regulating alloimmunization. Rather, they show that deletion of the C3 gene has effects. The absence of C3 may alter how the immune system develops or C3 may affect the expression of other genes that are directly involved in alloimmunization. However, the direct involvement of C3 is the simplest explanation. If C3 directly affects RBC alloimmunization then the precise mechanism by which C3 decreases the magnitude of RBC alloimmunization remains unclear but appears not to be through the CR1/CR2 complement receptors, since CR1/CR2 knockout mice have decreased and not increased alloimmunization. ²⁰ However, there are other complement receptors (i.e., CR3 and CR4) and the complement cascade activates numerous downstream immune modulatory molecules that depend upon C3 activation. Moreover, complement intracellular effects of C3 have also been described. ³⁵

In contrast to suppressing alloimmunization, deletion of CR1/CR2 has the same effect as deleting C3 regarding changing RBC alloimmunization to KEL‐K2_med from a CD4⁺ T cell independent to a CD4⁺ T cell dependent response. ²² Thus, CR1/CR2 is a likely candidate for C3 effects on CD4⁺ T cell dependence. Future studies will be required to elucidate the mechanisms by which deletion of the C3 gene has these effects.

It is worth noting that the above C3 effects have only been described with regard to the KEL‐K2_med RBC as an immunogen. To the best of our knowledge, no other RBC alloantigen induces a significant alloantibody response in the absence of recipient inflammation or is CD4⁺ T cell independent. The widely studied HOD antigen and human glycophorin A (hGPA) have little to no alloimmunization without recipient inflammation and both HOD and hGPA are CD4⁺ T cell dependent. The properties of KEL‐K2_med RBCs are not intrinsic to the KEL molecule since the same antigen at a higher copy number (KEL_Hi) has the same immunogenic properties as HOD and hGPA. Every transgenic mouse generated through random integration of an injected transgene risks confounders from inadvertent disruption of whatever gene may be present at the transgene integration site. However, we have carried out precise mapping of the KEL‐K2_med transgene integration site and no disruption of any genetic elements known to be involved in RBC biology or immune responses was observed. ³⁶ As such, the most likely interpretation is that there is a zone of alloantigen copy number that has the capacity to induce a CD4⁺ T cell independent alloantibody response in the absence of recipient inflammation.

It is unclear to what extent the unique properties of KEL‐K2_med RBCs model alloimmunization in humans. However, the copy number of blood group antigens in humans varies widely raising the distinct possibility that alloimmune responses to the same alloantigen may be CD4⁺ T cell dependent or independent based upon the copy number found in a particular donor. Importantly, the average copy number of RBC alloantigens on human RBCs correlates with estimated immunogenicity. ³⁷ It is also unclear if the effects observed for C3 in mice translate into humans and this will have to be tested. However, it is worth noting that increased rates of RBC alloimmunization are observed in inflammatory states where C3 is known to be consumed by underlying pathology, such as systemic lupus erythematosus (SLE) and in sickle cell disease (SCD). ³⁸ , ³⁹ , ⁴⁰ Of course, numerous other alterations are present in SLE and SCD, and the low C3 is just an association, but one that may require closer consideration given the murine data. Finally, the importance of elucidating mechanisms of alloimmunization to transfused RBCs notwithstanding, RBCs represent a unique immunogen that can help elucidate properties of immunology in general. As such, the findings by Mener et al. have broader potential relevance to understanding the nuances of immunology than just transfusion.

In aggregate, the current manuscript approaches a surprising and innovative finding with the level of reproducibility and rigor that are required to address known confounders in murine systems in light of surprising findings. Having controlled for common confounders, we conclude the presence of the C3 gene is important by whatever mechanism it exerts its effects. The novel discoveries by Mener et al., represent an innovative new area for which ongoing mechanistic research promises to be fruitful in understanding RBC alloimmunization and immunity in general.

CONFLICTS OF INTEREST

JCZ is a co‐founder of Svalinn Therapeutics and consults for Cerus Corporation.

Supporting information

FIGURE S1: Alloimmune response to KEL‐K2med is CD4+T cell dependent in Jackson C3KO B6 congenic mice (Jax 029661). Isotype or GK1.5 depleted B6 and Jackson C3KO B6 congenic mice (Jax 029661) were transfused with KEL‐K2med (See experimental scheme) and peripheral blood was collected at Days 7 and 14 post transfusion. The amount of anti‐KEL IgM at Day 7 and anti‐KEL Igs at day 14 were measured by a flow cytometry based cross match assay using either KEL‐K2_hi or B6 RBCs as targets as detailed in Section 2. Adjusted MFIs are calculated by subtracting the background antibody signal on antigen‐negative B6 RBCs from RBCs expressing KEL‐K2_hi. p values were calculated using a repeated‐measures, two‐way ANOVA with Šidák's multiple‐comparisons test and are designated as p > .05 (NS) and *p < .05, **p < .01, and ***p < .0005.

Jash A, Hay A, Zimring JC. Suppression of RBC alloimmunization and regulation of CD4⁺T cell dependence by C3 is not due to genetic confounders in mice. Transfusion. 2026;66(1):178–186. 10.1111/trf.70026

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.